A series of drug-like compounds derived from Sildenafil, Vardenafil and Tadalafil analogues were modelled through the MIA-QSAR (multivariate image analysis applied to quantitative structure-activity relationships) ligand-based approach. A highly predictive model was achieved and novel compounds, miscellany of substructures of these three representative phosphodiesterase type-5 (PDE-5) inhibitors were predicted using the calibration parameters obtained through partial least squares (PLS) regression. The high bioactivities of eight promising compounds were corroborated by docking evaluation. Calculated ADME-Tox (absorption, distribution, metabolism, excretion and toxicity) profiles for such compounds suggest advantages of some of them over the currently available, most common drugs used for the treatment of erectile dysfunction.